Trying to run Super Learner with a Pipeline throws a data.table error about sl3

[jeremyrcoyle]

Ran into the same error in a different context today. It looks like in your example, it's these lines that are causing data.table to get upset:
https://github.com/tlverse/sl3/blob/devel/R/sl3_Task.R#L105-L106

Commenting them out makes the error go away. I think those lines are a holdover from the interaction definition code you ported from gridisl. Do you know what they are meant to do?

[osofr]

Hm... It was obviously a hack for something. Most likely some kind of weird data.table behavior. Since we are modifying it by reference here, perhaps it would get upset when other variables are also pointing to the same data.table. Lets try removing it and see what happens?

[jeremyrcoyle]

The MRE works now. Closing this stale issue